On tis, 2007-09-18 at 18:53 +0200, Thomas-Martin Seck wrote:
> * Henrik Nordstrom ([EMAIL PROTECTED]):
>
> > It would be great if you could test the next 2.6 nightly snapshot. This
> > is quite likely identical to 2.6.STABLE17 unless something critical pops
> > up in the next days..
>
> I just di
* Henrik Nordstrom ([EMAIL PROTECTED]):
> It would be great if you could test the next 2.6 nightly snapshot. This
> is quite likely identical to 2.6.STABLE17 unless something critical pops
> up in the next days..
I just did a quick build test of the 20070918 snapshot and it builds ok
on FreeBSD 7
On mån, 2007-09-17 at 19:18 +0200, Thomas-Martin Seck wrote:
> The patch works, thanks (tested against STABLE16 with -O0 and -O2)!
>
> I filed the issue as bug #2086. The additional debug info from frame 1
> you requested is included in this bug report in case you still need it.
Thanks!
Will re
* Henrik Nordstrom ([EMAIL PROTECTED]):
> On sön, 2007-09-16 at 19:54 +0200, Henrik Nordstrom wrote:
> > On sön, 2007-09-16 at 13:24 +0200, Thomas-Martin Seck wrote:
> >
> > > The segfaults occur only on the HEAD FreeBSD branch (7-CURRENT). We
> > > tried to see whether this is a compiler issue (
On sön, 2007-09-16 at 13:24 +0200, Thomas-Martin Seck wrote:
> The segfaults occur only on the HEAD FreeBSD branch (7-CURRENT). We
> tried to see whether this is a compiler issue (FreeBSD 7 will ship with
> gcc 4.2.1 as the system compiler while FreeBSD-6 uses 3.4.6) and found
> that it probably i
On sön, 2007-09-16 at 19:54 +0200, Henrik Nordstrom wrote:
> On sön, 2007-09-16 at 13:24 +0200, Thomas-Martin Seck wrote:
>
> > The segfaults occur only on the HEAD FreeBSD branch (7-CURRENT). We
> > tried to see whether this is a compiler issue (FreeBSD 7 will ship with
> > gcc 4.2.1 as the syste
* Adrian Chadd ([EMAIL PROTECTED]):
> It wouldn't be the first time - Squid had to be compiled under -O0 for
> so very long because of a random crash in the Squid internal DNS code;
> the FreeBSD gcc was generating very very bogus argument handling code.
>
> Wonder if something similar is going o
It wouldn't be the first time - Squid had to be compiled under -O0 for
so very long because of a random crash in the Squid internal DNS code;
the FreeBSD gcc was generating very very bogus argument handling code.
Wonder if something similar is going on here; try compiling it with -O0?
(I'll read
Squid folks,
I am currently struggling with an odd issue with 2.6.STABLE16's cf_gen
that (unfortunately) keeps me (or rather the assigned FreeBSD committer)
from updating the FreeBSD port to 2.6.STABLE16. (In the meantime, we ship
STABLE15 plus all official changesets.)
The segfaults occur only o